基于多视角反射光谱的草地叶面积指数遥感反演

随着观测角度和观测平面的变化,同一地物产生的反射率、辐射信息也会产生很大差异。通过地面观测草地多平面多角度光谱反射率和叶面积指数数据,分析了光谱反射率随观测平面、观测角度的变化规律,并模拟TM资料红光(630～690 nm)、近红外(760～900nm)和蓝光(450～520 nm)波段反射率,计算RVI、DVI、NDVI、SAVI、ARVI和MSAVI等不同植被指数,分析不同观测平面和角度下植被指数与叶面积指数的相关性,建立遥感反演模型。结果表明,主平面上光谱反射率各向异性强于垂直主平面,且前向反射率低于后向,垂直主平面上前向和后向反射率呈现一定的对称性。主平面后向观测时植被指数与叶面积指数的相关性大于主平面前向观测、垂直主平面前向和后向观测的相关性。在建立的一元线性、乘幂、对数、指数、多项式等回归模型中,以三次多项式回归模型反演叶面积指数的精度最高。主平面上后向观测的NDVI、RVI、SAVI、ARVI的三次多项式回归模型均达到显著水平,在不同的观测角下,以-30°、-45°时拟合精度最高。

Study on the Estimation of Vegetation Leaf Area Index Based on Reflectance Spectra in Different Directions

With the change of view angle and plane,the same surface produces different reflectance and radiation.The change rule of spectral reflectance at different view angles and planes is analyzed based on ground observation of grass spectral reflectance and LAI data.We simulated the reflectance of red,NIR and blue bands of Land sat TM,and then calculated vegetation index such as RVI,DVI,NDVI,SAVI,ARVI and MSAVI.The correlation of vegetation index and LAI at different view angles and planes was analyzed and the regression models for retrieving LAI were constructed.The results showed that the anisotropy characteristics of spectral reflectance factors at solar principal plane were stronger than that at orthogonal principal plane.The spectral reflectance at backward direction was larger than that at forward direction of solar principal plane and they were symmetrical at orthogonal principal plane.The correlation of vegetation index and LAI at backward direction of solar principal plane was higher than that at forward direction of solar principal plane and orthogonal principal plane.The cubic polynomial regression model was the best in the different forms of linear,logarithmic,exponential,exponentiation and polynomial regression models.At backward direction of solar principal plane,the cubic polynomial regression models of NDVI,RVI,SAVI and ARVI have all reached significant levels,especially at the view angles of-30° and-45°.